摘要
为提高粮食产量的预测精度,合理预测其发展趋势,保障粮食安全,选取2000-2021年中国粮食产量、农业总产值和农业就业人员数据,构建了ARIMA,GM,VAR,MGM,ARIMA-GM-BP,VAR-GM-BP,ARIMA-MGM-BP和VAR-MGM-BP等一系列模型,并以MAPE为评价指标对模型进行拟合和预测精度比较.结果表明:多变量预测模型在预测精度方面优于单变量预测模型,而多变量组合模型又优于单变量组合模型;进一步表明,提升组合模型中单一模型的预测精度有助于提升组合模型的预测精度.最后,研究构建的VAR-MGM-BP组合模型拥有最小的MAPE值,并利用VAR-MGM-BP模型对中国未来五年的粮食产量进行预测.
In order to improve the prediction accuracy of grain output,reasonably predict its development trend and guarantee food security.A series of models such as ARIMA,GM,VAR,MGM,ARIMA-GM-BP,VAR-GM-BP,ARIMA-MGM-BP,ARIMA-MGM-BP and VAR-MGM-BP are constructed by selecting the data of China's grain output,agricultural gross output value and agricultural employed persons from 2000 to 2021,and MAPE is used as the evaluation index for model fitting and forecasting accuracy comparison.The results show that the multivariate prediction model is better than the univariate prediction model in terms of prediction accuracy,and the multivariate combination model is in turn better than the univariate combination model,it is further shown that improving the prediction accuracy of a single model in the combination model helps to improve the prediction accuracy of the combination model.Finally,the VAR-MGM-BP combined model constructed in this study has the smallest MAPE value,and the VAR-MGM-BP model is utilized to predict China's grain output in the next five years.
作者
王夏鑫
周健
WANG Xia-xin;ZHOU Jian(School of Mathematics and Statistics,Kashgar University,Kashgar 844000,Xinjiang,China;School of Economics and Management,Qilu Normal University,Jinan 250202,China)
出处
《兰州文理学院学报（自然科学版）》
2024年第3期30-38,共9页
Journal of Lanzhou University of Arts and Science(Natural Sciences)
